Fast 3D Thinning of Medical Image Data based on Local Neighborhood Lookups

Abstract

Three-dimensional thinning is an important task in medical image processing when performing quantitative analysis on structures, such as bones and vessels. For researchers of this domain a fast, robust and easy to access implementation is required. The Insight Segmentation and Registration Toolkit (ITK) is often used in medical image processing and visualization as it offers a wide range of ready to use algorithms. Unfortunately, its thinning implementation is computationally expensive and can introduce errors in the thinning process. This paper presents an implementation that is ready to use for thinning of medical image data. The implemented algorithm evaluates a moving local neighborhood window to find deletable voxels in the medical image. To reduce the computational effort, all possible combinations of a local neighborhood are stored in a precomputed lookup table. To show the effectiveness of this approach, the presented implementation is compared to the performance of the ITK library.